U.S. President Donald Trump has signed an executive order titled “Preventing Woke AI in the Federal Government,” barring federal agencies from buying artificial-intelligence systems that the White House deems ideologically biased. The measure, unveiled on 23 July alongside a broader AI Action Plan, instructs agencies to procure only large language models that meet two principles—“truth-seeking” and “ideological neutrality”—and explicitly denounces content related to diversity, equity and inclusion, critical race theory and similar themes. The order gives the Office of Management and Budget, the General Services Administration, the Office of Federal Procurement Policy and the Office of Science and Technology Policy 120 days to issue implementation guidance. Each federal agency must then adopt procedures within 90 days to ensure any AI system it acquires complies with the new standards. Technology companies with lucrative government contracts could face significant technical and commercial hurdles. Researchers say achieving politically neutral outputs is technically unproven, and modifying model behaviour to match a single administration’s worldview risks alienating global users and conflicting with existing commitments to accuracy and fairness. Amazon, Google, Microsoft, OpenAI, Anthropic and Elon Musk’s xAI all provide AI or cloud services to U.S. agencies and could be required to re-evaluate training data and safety filters to preserve federal business. The policy lands just days after the Pentagon awarded contracts worth up to $200 million each to Anthropic, Google, OpenAI and xAI to develop AI tools for national-security tasks, underscoring both the government’s dependence on frontier models and the uncertainty over how the new ideological standards will be enforced.